Skip to main content

Full text of "Developing and using profiles : a primer for PSRO physicians"

mi7 



DEVELOPING AND USING PROFILES: 
A PRIMER FOR PSRO PHYSICIANS 




October 19, 1978 



MANUALS 
RA 
399 
A3 

TA-17 
1978 



U.S. DEPARTMENTof HEALTH, EDUCATION, and WELFARE 

HEALTH CARE FINANCING ADMINISTRATION 
HEALTH STANDARDS AND QUALITY BUREAU 



J 



J 



M 



MEMORANDUM^ 



TO 



Planning and Conditional PSROs; 
Statewide Councils; Regional 
PSRO Project Officers 



DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE 
HEALTH CARE FINANCING ADMINISTRATION 
HEALTH STANDARDS AND QUALITY BUREAU 
OFFICE OF PROFESSIONAL STANDARDS REVIEW ORGANIZATIONS 



date: OCT 1 9 1978 

TECHNICAL ASSISTANCE 
DOCUMENT NO. 17 



FROM 



Director 



subject: "Developing and Using Profiles; A Primer for PSRO Physicians" 

"Developing and Using Profiles; A Primer for PSRO Physicians" was 
developed by InterQual, Inc., under their expert assistance contract 
(No. 240760106) with the Health Standards and Quality Bureau (HSQB) . It 
is intended to assist PSROs in their efforts to provide PSRO physicians 
with an understanding of the concept of profiles and profile analysis. 



or 



Section 1155(a)(4) states that each PSRO "shall be responsible for 
arranging for the maintenance of and the regular review of profiles 
care and service received and provided with respect to patients. 
Profiles shall also be regularly reviewed on an ongoing basis with 
respect to each health care practitioner and provider to determine 
whether the care and services ordered or rendered are consistent with 
criteria specified..." in other sections of the law. Section 101.711 of 
the proposed rules on "Procedures for Review of Hospital Services" 
reiterates the requirement that PSROs perform profile analysis on the 
health care provided to Federal patients by health care practitioners 
and providers. 

While PSROs are aware of the charge that they develop profiles and 
utilize profile analysis as part of their review activities, they often 
experience difficulty in explaining profiles and profile analysis to 
PSRO physicians and staff who are not familiar with these concepts. It 
is for this reason that we requested InterQual, Inc., to develop this 
primer. It is designed to make available to PSROs documentation which 
can help them provide PSRO physicians and staff with a working knowledge 
of profiles and profile analysis. It is intended that this knowledge 
will enable the PSRO physicians to utilize the results 



PSRO Reference: Transmittal Number 61 

PSRO Program Manual Chapter: 

PSRO Technical Assistance Document No: 



/ 



Technical Assistance Document No. 17 



Page 2 



of the PSRO's profiling efforts to set meaningful objectives for the 
PSRO. Understanding profiles and profile analysis should also enable 
PSRO physicians to analyze the impact of PSRO review activities on the 
quality of care and utilization of services in the PSRO area. 

We bel ieve that PSROs will find this primer useful. Please direct any 
comments or questions to your Regional Project Officer. 




Michael J. Goran, M.D. 



Attachment 



DEVELOPING and USING PROFILES 
A PRIMER 

FOR 

PSRO PHYSICIANS 



Developed by Martin L. Waldman, M.D., Director of Research and Development 
InterQual, Incorporated, 7^0 N. Rush St., Chicago, Illinois, 6061 1 . This 
work was performed pursuant to DHEW Contract No. 240-76-0160, for the Office 
of Professional Standards Review, Health Standards and Quality Bureau. Health 
Care Financing Administration. 



A PROFILE IS SIMPLY AN OVERVIEW 



1 profile Vpr6-ifil\ n [It profilo, fr. prvfilare to draw in outline, fr. 
pro- forward (fr. L) + filare to spin, fr. LL — more at file] 1 
: a representation of something in outline; esp : a human head or 
face represented or seen in a side view 2 : an outline seen or 
r\ r\ n\ir\ represented in sharp relief : contour 3 : a side or sectional 

VJPPCTpp QflYS elevation: as a : a drawing showing a vertical section of the 

nLUO I L l\ On 1 O ground b : a vertical section of a soil exposing its various zones 

or inclusions 4: a set of data often in graphic form portraying the 
significant features of something <a corporation's earnings ~->; esp 
: a graph representing the extent to which an individual exhibits 
i raits or abilities as determined by tests or ratings 6 : a concise 
biographical sketch ayn see outline 

'profile vt pro-filed: pro fil ing 1 : to represent in profile or by 
a profile : produce (as by drawing, writing, or graphing) a profile 
of 2 : to shape the outline of by passing a cutter, around — 



THE SKYLINE IS A PROFILE 
OF THE CITY 




BALANCING A CHECKBOOK - THE COMPARISON OF TWO PROFILES 



t ■( SU« 10 OtPHC *N» PM QW CHAHGIS 0" 



Jiff 
trio 



% ; Cherry Kill 
*fi\ Central Cghf^Fe ait- 




Seaside Company 
10 East 1st Street 



Statement of account wiek j 

the sscl p ity UAtm 



Date 


Check, and Other Debits 


Deposit, ■ 


MilM 


Oct. 1 


Balance brought forward 




2.!«3.00 


2 


24O0O 




113)03.00 


3 


205.00 


315.00 


2.0I3.M 


5 


175.00 


29500 


2.IJ3JW) 


6 


310.00 




1,»23.<X> 


12 


190.00 


425.00 


2,0*^.01 


!5 


135.00 255.00 


235.00 


1. "03.00 


22 


260.00 


535.00 


2.17*00 


21 


210.00 280.00 


113.00 


I , m ?x™ 


JO 




550.00 


2.353.CT 


31 


1 15.00 NSF 5.00 SC 


39* t") CM 


2.*3t.m 



CM Credit memorandum 
DM Debit Memorandum 



NSF Not sufWnem fu«*** cV^S: 
SC Servi« charge 



The various stockmarket indexes (Dow- Jones, Standard and Poors, etc.) provide 
A PROFILE OF THE ECONOMY. A PSRO profile provides an overview of the informa- 
tion collected about medical care delivered within a PSRO area. 



2 



PSRO PROFILES CAN BE USED TO 



ESTABLISH REVIEW PRIORITIES FOR MCE 

- by identifying both problem areas and areas of strength 
and good performance 

FOCUS CONCURRENT REVIEW 

- by identifying those areas where 100% concurrent review 
is no longer needed (focus out) 

- by indicating areas where review should be intensified 
(focus in) 

MONITOR PERFORMANCE 

- by periodic profiling of key indicators where previously 
identified deficiencies have been corrected 

SURVEY CARE OF ALL COVERED PATIENTS 

- by periodic examination of key outcome indicators 

DEMONSTRATE IMPACT OF REVIEW 

- by displaying the changes that have occurred 

EXAMINE OTHER ISSUES 

- i.e., issues not related to review but which can affect the 
functioning of the PSRO such as characteristics of the com- 
munity population, reliability of data collection, etc. 




3 



ADVANTAGES OF PROFILES 



PROVIDE BROAD PROSPECTIVE 

By using aggregated data, profiles can indicate problem areas that 
are not visible when reviewing care on a record by record basis, 
i.e., while the care of each patient may seem appropriate, the 
care when viewed as a conglomerate and compared to other refer- 
ents seems to be out of line. 

ACCOUNT FOR EXPECTED VARIATION 

By displaying a range of values (spread) for a patient group, 
profiles can be used to evaluate whether the expected (desired) 
spread exists. For example, the shortest and the longest length 
of stay, or the most frequent and least frequent procedures. 



SHOW TRENDS 

Profiles that display the same measurements over time allow for 
analysis of trends that can help predict whether the current sta- 
tus is being maintained, slipping, or starting to improve. 



FORM OBJECTIVE BASIS FOR DECISION MAKING 

Based on objective observable features of patients/providers/ 
institutions, profiles can be used to document and support t'a?, 
basis for corrective action, thus removing any hint of partiality 
or capr iciousness. 



LIMITATIONS OF PROFILES 



The use of aggregated data describing groups of patients (thus concealing 
the identity and characteristics of the individual patient) imposes certain 
limitations that must be recognized if appropriate analysis is to be made. 
These limitations include: 

• Only comparative measurements are displayed - 
absolute judgments can not be made. 



• Patterns are displayed, not specifics - 
conclusions can not be drawn about care of 
an individual patient. 

• A profile as an overview requires analysis 
and interpretation - the raw data can often 
be misleading. 



The limitations of PSRO profiles can be overcorr*; - 
and profiles used to their fullest advantage - 
if analysis takes place as close as possible to 
the "point of service", that is, in the place where 
care is provided. Though profiles may be generated 
at a distance, valid analysis, including develops":/ 
additional information and providing explanations, 
can best be done by those who understand how local 
factors influence the data portrayed in the profile 

For example, an apparent abuse of acute care beds 
for Medicare patients with hip fractures could be 
due to a shortage of ECF beds in the area - but 
only someone with an understanding of looal prob- 
lems could offer that explanation. 



LOCAL ANALYSIS OF 



PROFILES 



5 



THE COMPUTER AND PSRO PROFILE ANALYSIS 



Many PSROs have not yet involved themselves in profile analysis because 
of a belief that this type of review is dependent on having a completely 
debugged computer system before profiles can be produced. 

While it is true that a computer is useful in storing large quantities of 
data, sorting the data in multiple ways, and making multiple and highly 
sophisticated correlations, this may not always be immediately available 
S.nce one of the basic characteristics of a useful profile is its simplic- 
ity, a reasonable approach is to design an initially simple profile (by 
excerpting existing data) that can be produced manually. If after test- 
ing the usefulness of the design through analysis, it is decided to use 
that format in an on-going monitoring situation or to apply it to multiple 
groups, then it may become worthwhile to incorporate that into the desiqn 
of the computerized data system. 

All of the computerized discharge abstract systems, some government agen- 
cies, and many of the fiscal intermediaries already have "canned" programs 
that produce displays resembling profiles. But all too often these dis- 
plays are lengthy, crowded, and so complex that it takes an expert to 
excerpt them before they become useful -- for otherwise an "information 
overload" occurs leading to confusion and misinterpretation. Such displays 
are best used as sources of data from which a member of the PSRO staff (who 
is knowledgeable in the organization and contents of the print-out) excerpts 
the pertinent data and displays it in a simple format, producing where nec- 
essary multiple simple profiles, rather than a complex, confusing concatena- 
tion of cascaded columns. 



6 



CONFIDENTIALITY AND PSRO PROFILES 



The issue of confidentiality has been raised as a barrier to producing PSRO 
profiles. The real issue is not whether those to whom data is trusted will 
give it away (break the confidentiality), but rather, what "security" is 
provided so that data will not fall into the hands of those who are not 
authorized to see it. 

The physician, in treating his patients, acquires information that he can 
not be forced to reveal. The medical record in the hospital is kept secured 
(in the medical record department) and only those with proper authorization 
are allowed access to it. 

Data abstracted from the record for PSRO purposes, is kept secured in a num- 
ber of different ways. First of all, regulations specify what data is to be 
abstracted and what portion of that data may be forwarded to other users, 
thus limiting the danger of theft and misuse. For example, patient and phy- 
sician identifiers are collected and retained at the PSRO - they are not 
transmitted to HSQB or DHEW. 

Secondly, through the aggregation of data for profiles, the details of the 
case history of an individual patient become invisible. Since aggregated 
data displays only patterns, no inference can validly be drawn about the 
care given to one patient nor can such data validly be applied as a standard 
to any individual case. Computerized data is further guarded in ways designed 
to limit access to those who need to know. Many data processing organizations 
require security oaths of their employees and have contractual agreements with 
their clients concerning what data may be released, to whom, and under what 
ci rcumstances. 

Access is further limited by the use of protective computer programming so 
that only those who are so authorized can retrieve information. Additional 
protection is provided through the use of confidential code numbers to record 
patient and provider identity, with the translation lists remaining at the 
local level. The PSRO should assure itself that such security precautions 
are taken by whoever handles data from their area. 

The other side of the issue of confidentiality is a matter of protection 
provided by policy. As mentioned before, HSQB policies keep patient and 
physician identities sequestered at the PSRO level, thus ensuring the con- 
fidentiality of data down to that level. The aggregated data displayed in 
profiles is by policy available to the federal government. Since the data 
is thus available, the best protection against improper use of the data 
lies in local interpretation rather than transmittal of raw data to be 
interpreted in a vacuum. That is, the analysis of the meaning of such data 
should be carried out by those who are familiar with local factors that may 
explain apparent deviations from expected practices. 



7 



THE ANATOMY OF A PSRO PROFILE 
A USEFUL PSRO PROFILE 



IS SIMPLE 

Relationships and differences are much clearer if a profile 
is concerned with only one thing at a time. Multiple pro- 
files each displaying only one data item at a time are eas- 
ier to analyze than a complex profile with many entries. 



HAS A DEFINED PURPOSE 

Defining the purpose leads to logical design. Creating a 
profile with no purpose in mind is a waste of time and energy 
(both yours and others) and can lead only to confusion and 
mi s i nterpretat ion. 



DISPLAYS AGGREGATED DATA 

Aggregating data discloses patterns of care that cannot be 
perceived when looking at one case at a time, thus revealing 
problems that may otherwise remain hidden. 



8 



PROVIDES REFERENCE POINTS 



Reference points provide the basis for comparison that is 
essential in analyzing a profile. Descriptions of perfor- 
mance alone are meaningless without something against which 
they can be evaluated. 



YIELDS NEW INFORMATION 

The comparison of performance patterns and reference points 
reveal differences and similarities from which inferences 
may be drawn and predictions made. These findings and con- 
clusions form the basis for decisions regarding additional 
act i vi ty . 



LEADS TO RECOMMENDATIONS 

Recommendations for further activity which may include gath- 
ering of additional information, modifying review plans, insti 
tuting corrective action, etc., can thus be based on and sup- 
ported by objective information - intuition isn't good enough 
anymore. 



9 



THE COMPONENTS OF A PSRO PROFILE 



THE GROUP of patients 
covered by the profile 



SUB-GROUPS into which 
the group is divided 



SELECTEO DETAILED 
DIAGNOSIS GROUP 
( 1 ) 



MEASUREMENTS of 
current comparison 



REFERENCE POINTS 
for comparison 



STAY PROFILE FOR 18 SELECTED DIAGNOSIS GROUPS 
NATION. NnpTHFAS'FRN REGION. AND PSRO 3 



JANUARY - JUNE 197/ 



-ALL TITLES- 
TOTAL 
PATIENTS-1 

(2> 



DAYS 

10 11 12 IS 



DIS OF GB I BILE DUCT WITH OP. W/O 2ND DX 

NATION 4006 

REGION 1565 

PSRO 86 

DIS PROSTATE. W/TRANSUR PROST, W/O 2ND DX 

NATION 4674 

REGION 130*. 

PSRO 5* 

DIS OF FEM GEN ORG W/HY5 T ER EC . T H MA J OP 

NATION 7770 

REGION 2708 

PSRO 2*5 

NORMAL DELIVERY. W/O OP OR W/MINOR OP 

NATION 29704 

REGION 9694 

FSRO 1626 

ARTHRITIS W/REPAIR. PLASTIC OP . L AMNECTOMY 

NATION 4327 

REGION 1046 

PSRO 43 

FRACTURE OF F EMUR .PELVIS. MULTIPLE W/O OP 

NATION 5310 

REGION 1469 

PSRO 127 




NATION / N- 

REGION 

PSRO /+ P ♦ 

2|th to 75th Percentile, first and last » 
mirk end poi/ts. letter marks median) 



TiY-2 

4 115 16 17/ 18 19 20 21 22 23 24 25 >25 



87 



2-EXCLUDE5 DEATHS ! IF 75TH PERCENTILE EXCEEDS 25 DAYS . IT WILL BE SHOWN AS >25 DAYS . 



10 



TYPES OF PSRO PROFILES 



The type of PSRO profile is determined by how the GROUP is selected, 
the SUB-GROUPS follow therefrom: 



TYPE OF PROFILE 


GROUP 


SUB-GROUPS 


Patient 


All patients in the PSRO area 
defined by patient character- 
istics such as diagnosis/prob- 
lem/condition, type of therapy, 
age group, etc. or tracking 
individual patients, if the 
review system includes hospital, 
ambulatory and long term care 
sett i ngs. 


By Hospital 

By CI inical Service 

By Provider, etc. 


Provider 


Al 1 patients in the PSRO 
defined by who provided the 
service such as patients of 
Dr. X, patients under care 
of anesthesia group, patients 
whose x-rays were read by 
Dr. Y. , etc. 


By Diagnosis/ 

Problem/ 

Cond i t ion 
By Type of Therapy 
By Age Group 
By Hospital 


Inst i tut ion 


Patients treated in one hos- 
pital or in an internal divi- 
sion of a hospital (service, 
special care unit, etc.) 


By Provider 

By Diagnosis 

By Type of Therapy 

By Age Group, etc. 



11 



THE GROUP 



A clear definition of the group of patients included in a profile will 
conta in : 

1. The characteristic on which the patients are selected 

e.g., diagnosis of acute myocardial infarction, 
patients undergoing cholecystectomy, patients 
treated in Hospital A, patients treated by Dr. X, 
etc . 

2. The universe from which the group was selected 

e.g., all federally financed patients in PSRO 
Area X. 

3. The time period covered 

e.g., patients discharged during Jan-March 1 977 • 



GROUP 

1. CHARACTERISTIC 

2. UNIVERSE 

3. TIME PER I C 




STAY PROFILE FOR 18 SELECTED DIAGNOSIS GROUPS 
■•NATION, NORTHEASTERN REGION , AND PSRO B 

JANUARY - JUNE 197/ 
. -ALL TITLES- 

SELECTED DETAILED „.J?ISt e i l ~~3> J.~~l~~s"2~~l 8 9 lo"ll~12 13 14 15 ] 

DIAGNOSIS GROUP PATIENTS-1 123^56789 10 " " "J 

DIS OF GB t BILE DUCT WITH OP, W/0 2ND DX 

NATION 40 06 + N + 

REGION 1565 + R — + 

PSRO 86 + P - 1 

DIS PROSTATE. W/TRANSUR PROST, W/0 2ND DX 

NATION 467* +--N + 

REGION 1304 +--R + 

PSRO 54 + P + 

DIS OF FEM GEN ORG W/HYST EREC , OTH MA J OP 

NATION 7770 + N + 

REGION 2708 +--R + 

PSRO 243 + P + 

NORMAL DELIVERY, W/0 OP OR W/MINOR OP 

NATION 29704 ♦ — M--+ 

REGION 9694 + --R + 

FSRO 1626 +P--+ 



12 



SUB-GROUPS 



The GROUP is divided into SUB-GROUPS so that performance may be compared. 

The display may clearly indicate: 

1. The characteristic used to establish the SUB-GROUPS 

e.g., by hospital, by provider, by diagnosis, 
by age group, etc. 



NOTE: SUB-GROUPS must be chosen so that every 
patient in the group fits into one and 
only one SUB-GROUP. 



2. The factor that determines the order in which the 
SUB-GROUPS are listed (since this may be very sig- 
nificant) 

e.g., ranked by size, ranked by performance, 
listed in diagnosis code number order, etc. 



GROUP 

1. CHARACTER I ST I 

2. UNIVERSE 

3. TIME PERIC 



SUB-GROUPS 

1 . By Diagnos is 
Group 




Numer i c 
Order 
Di agnos i s 
Code ■ 



T AY PROFILE FOR 18 5ELECTED DIAGNOSIS GROUPS 
A T ION . NnPTMfAS'ERN REGION. AND PSRO 3 
*^ 

^ JANUARY - JUNE 197/ 



SELECTED DETAILED 
DIAGNOSIS GROUP 



ALL TITLES- 
TOTAL -- DAYS STAY-2 ! 

PATIENTS-1 1 2 3 4 5 6 7 8 9 10 11 12 1} 1* 15 16 17 IS ! 



DIS OF CP I BILE DUCT WITH OP.W/0 2ND DX 

NATION 4006 

REGION 1565 

PSRO S6 

DIS PROSTATE. U/TRAN5UR PROS T , Kits 2ND DX 

NATION 6676 

REGION 1304 

PSRO 5* 

DIS OF FEP1 GEN ORG U/HYSTEREC . IH MA J OP 

NATION 7770 

REGION 2708 

P5R0 243 

NORMAL DELIVERY. U/0 OP OR U/fllNOR OP 

NATION 29704 

REGION 9694 

PSRO 1626 

ARTHRITIS WV REPAIR, PLASTIC OP . L AMNECTOMY 

NATION 4327 

REGION 1046 

PSRO 43 

FRACTURE OF FEMUR . PELVIS. MUl MPLE U/O OP 

NATION 5310 

REGION 1469 

PSRO 127 



-N--» 
♦ --R» 
»P--» 



13 



THE MEASUREMENTS 



A measurement consists of 
an ELEMENT 



pi us 



one or more UNITS 



pi us 



the VALUE (S) 



the thing being measured e.g., 
LOS, incidence of occurrence, 
presence of variation from 
criteria, age, etc. 



of measurement such as total 
number, percentages, averages 
percent! les , etc. 

for the GROUP, each SUB-GROUP 
and each REFERENCE POINT ex- 
pressed in terms of the units 



To understand the significance of a measurement, it is helpful 
to dissect it into its component parts. In the example promise 
on the facing page, there are two different measurements displayed 



SIZE of the SUB-GROUPS 

ELEMENT: Patients in selected detailed diagnosis group 

UNIT: Total number of such patients 

VALUES: Displayed as numbers for each diagnosis group 

EXAMPIE: Tht PSRO had %b patlznU vtiXh dU- 
(loaza oh the. gcJU hladdoji and bile 
duct duAlng lom.afi.ij - June 1977, 
tiiz titQlon had '1565, and ike natX.on 
had 4006. 

STAY PATTERN 

ELEMENT: Days of stay displayed as a graphic scale across 
the page 

UNIT: 3 units are used; the 25th percentile, the 50th 

percentile, and the 75th percentile. 

VALUES: Graphically displayed as points along the scale 
for the PSRO, the region and the nation (sec 
key in upper right-hand corner) 



111 



EXAMPLE: 



PIS EASES OF GALL BLAWER 
ANV BILE VUCT 



PSRO 



REGION NATION 



25th peActntile. 
50 th p<tAc<LvvtLlz 
75th ptuctntiZt 



11 
16 



10 
15 



8 
10 
14 



SIZE of GROUP 
and SUB-GROUPS 



UNIT 



ELEMENT 



VALUES 



STAY PATTERN 
ELEMENT 



UNITS 



VALUES 




SELECTED DETAILED 
DIAGNOSIS GROUP 
( 1 ) 



DIS OF G? I BILE DUCI WITH OP.W/O 2ND DX 

NATION ^4006 

REGION ~1565 

PSRO 86 

DIS PROSTATE. W/TRANSUR PROST. WO 2ND DX 

NATION 4674 

REGION 1304 

PSRO 54 

DIS OF FErt GEN ORG U/HY5T EREC . OT H MA J OP 

NATION 7770 

REGION 2708 

PSRO 243 

NORMAL DELIVERY. U/0 OP OR U/MINOR OP 

NATION 29704 

REGION 9694 

FSRO 1626 

ARTHRITIS U/REPAIR. PLASTIC OP . L AflNEC TOMY 

NATION 4327 

REGION 1046 

PSRO 43 

FRACTURE OF FEMUR. PELVIS. MULTIPLE U/O OP 

NATION 3310 

REGION 1469 

PSRO 127 



♦ --R» 

♦ P--* 



^"ciUOES ollwl. IF 75TH PERCENTILE EXCEEDS 25 DAYS. IT WILL BE SHOWN AS >25 DAYS. 



15 



WHAT IS BEING MEASURED? 



Different units describe different characteristics of a group. Inferences 
and conclusions can only be appropriately drawn when there is a clear under- 
standing of what is being measured - knowing that someone weighs 170 pounds 
does not tell how tall he is, nor does it tell whether he is "fat" or "skinny." 
Commonly used units are displayed below, along with an explanation of what 
characteristics each measure. Examples are given along with the profile on 



the facing page. 


UNIT 


CHARACTERISTIC 


Total Number 


Size of group and sub- group 


Average 
(Mean) 


An estimate of the central tendency of the 
values. It has the basic defect of being 
contaminated by outliers. 






Standard Deviation 
(SD) and Coefficient 
of Variation 


The standard deviation measures the absolute 
variation about the mean. The coefficient 
of variation is the standard deviation divided 
by the mean, which is a more relative measure 
and thus more useful in analyzing length-cf- 
stay data. In a reasonably homogeneous g rous- 
ing one might expect the coefficient of varia- 
tion to be approximately .5* A lower value 
indicates a tighter stay distribution, and s 
higher value reflects less consensus on the 
appropriate length of stay and treatment pat- 
tern for the group of patients studied or less 
homogeneous In the group. 


Percent i les 


■ i 

Statements of what value various prooor t ; ens 
of the group have attained, e.g., the 10th 
percentile stay is that day on which 10'n of 
the patients have been discharged. A per- 
centile pattern is a description of the spread 
of the group from a starting point - usual ?y 
the lowest value. 


Med ian 


1 - " — ~1 

The median is the 50th percentile, and Is also 
an estimate of the central tendency of the 
group. A comparison of the average and the 
median will indicate how close the distribution 
approaches a normal one and in which direction 
it is skewed, i.e., shifted away from or ' ng 
symmetrical. The median overcomes the defect 
of contamination of the mean by the outliers. 





16 



UTILIZATION IV SCIfien fllA'LEI) DIAGNOSIS CROUPS 
JAHUAi JUNE 197? 



SERIES 2 
TABLE 1 



- - ALL TITLE DISCHAPCES -- 

AVERAGE STAY PERCENTILE STAY IN DAY5-4 

SELFCICI) DETAILED DIAGNOSIS GPOUP TniAL NO. Or TOTAL DFATHS LONG STAY STUDIED AVG. COEFF 10TH 25IM 50TH 75TH 90TH 



DISCHARGES 
< 1 ) (?) 


DAYS 
( J ) 


(47 


PTS.-l 
(51 


PIS. -2 
(e) 


STAY 
( 7 ) 


VAR-3 
(8) (9) 


(10) 


(11) 


< 12 ) 


( 13 ) 


DIABETES. AGE -35 W/O OP U/O 2ND DX 


154 


19SJ 


6 


13 


135 


10 


5 





5 4 


7 


10 


15 


27 


NEUROSES .PERSONAL ITY DISORDERS 


74 


1 362 





1 


73 


17 


.6 





7 4 


8 


16 


24 


3 7 


DIS OF LYE WITH EXTRACTION OF irNS 


57« 


3702 


2 


23 


551 


5 


e 





4 4 


5 


5 


7 


1 


ACUIE MYOCARDIAL INIARC1I0N 


:ss 


1006S 


1 9 


1 7 


462 


17 


■? 





5 7 


10 


16 


24 


33 


ISCHEMIC HEART DIS, W/O OP W/O 2ND DX 


m 


1155 


1 


5, 


1 10 


7 


3 





6 3 


5 


8 


1 


20 


CEREB THROMBO-EMBOl W/O OP. W/O 2ND DX 


101 


1*30 


1 5 


9 


77 


I 3 


4 





5 5 


7 


14 


20 


4 


HYPERTROPHY OF T i A 


255 


432 





4 


251 


1 


6 





4 1 


1 




2 


2 


ACUTE UPPER RE5P. INF.IINFLU. AGE>44 


29 


295 





1 




? 


6 





7 3 


4 


7 


13 


24 


PNEUMONIA. AGE-:31 


210 


1345 


3 


6 


201 


3 


a 





5 3 


4 


5 


8 


11 


PNEUMONIA. AOEMO.W/O OP W/O 2ND DX 


92 


480 


5 


2 




8 


f. 





5 3 


5 




1 3 


1 9 


GASTRIC > PEPTIC ULCER W/O OP W/O 2ND DX 


35 


270 


1 


1 


33 


7 


1 





4 3 


5 


7 


9 


12 


HERNIA OF ABD CAV, AGE '6 A W/MINOR REPAIR 


252 


247 1 


2 


12 


2 38 


8 


5 





5 4 


6 


8 


11 


17 


DIS OF GB t BILE DUCT WITH OP. W/O 2ND DX 


86 


1146 





6 


80 


11 


.» 





4 7 


8 


11 


16 


23 


DIS PROSTATE. W/TRANSUR PROST.W/O 2ND DX 


59 


621 








54 


1 1 


.5 





4 6 


9 


11 


14 


16 


DIS OF F EM GEN ORG W/HYSTEREC . TH MA J OP 


294 


2106 


1 


7 


251, 


7 


8 





5 4 


5 


7 


10 


14 


NORMAL DELIVERY, W/O OP OR W/MINOR OP 


1626 


5657 





19 


1607 


3 


4 





3 3 


3 


3 


4 


5 


ARTHRITIS W/ REPAIR. PLASTIC OP . L AMNECTOMY 


43 


156 5 





7 


36 


29 


5 





3 16 


25 


30 


43 


69 


FRACTURE OF FEMUR. PELVIS. MULTIPLE W/O OP 


132 


3644 


3 


13 


114 


21 


fl 





7 4 


11 


21 


42 


61 


1-LIVE DISCHARGES WITH UNUSUALLY LONG STAYS. WHO WOULD 

DT P E NOEtt I UPON THE PATIENT'S PRINCIPAL DIAGNOSIS. 
J-EXCLUDtS DEMHS AND LONG STAYSUOLS. 4 AND 5). THESE 


SKEW THE 
PATIENTS 


AVERAGE 
ARE ALSO 


STAY. THE 
EXCLUDED 


DEFINITION 
FROM COLS. 


or 

7 


A 

kKD 


LONG STAY I 
8. 


S 









.3-THE STANDARD DEVIAIION DIVIDED BY THE AVERAGE STAY 
-4-EXCLUDES DEATHS ONLY. 



EXAMPLES (from excerpted profile above) 


Total number 


Column 2 - 15^t patients with diabetes were discharged during 
Jan-June 1977. 


Average 


Column 7 _ The average stay for diabetics was 10.5 days - note 
that this average is based on 135 patients (Col 6) - 
see footnotes to profile. 


Coefficient 
Of Variation 


Column 8 - The Coefficient of Variation for Hypertrophy of T&A 
(O.k) indicates a very small amount of dispersion, 
especially when compared to that for Neurosis 
Personality Disorders (0.7). 


Percent i 1 es 


Columns 9 - Of the diabetics at least 10% will have been dis- 
through 13 charged by the ktb day, 25% or more by the 7th day, 
50% or more by the 10th day, 75% or more by the 15th 
day, and 30% or more by the 25th day. 


Med ian 


Column 11 - The median for acute upper respiratory infection is 
7 days - since the average stay is larger (Col 7 - 9 • 6 
days) this means that the distribution is skewed to 
the right, i.e., toward longer stay. (Where the 
median is larger than the average as in Hypertrophy 
of T&A (see bottom line - Column 7-1.6, Col 11-2) 
the skew is to the left, i.e., toward shorter stay. 



17 



WHAT DATA IS AVAILABLE FOR PSRO PROFILES? 



Under federal programs two sets of data are collected on each patient. 
They are: ]) The Uniform Hospital Discharge 

Data Set (UHDDS) 
2) PSRO Hospital Discharge 

Data Set (PHDDS) 



The Uniform Hospital Discharge Data Set (UHDDS) was developed by the National 
Center for Health Statistics in 1971. The purpose for the development of the 
UHDDS was to have a minimum set of data, uniformly defined, capable of pro- 
viding basic and comparable information on all hsopital discharges to all users. 

Hospitals are encouraged to collect UHDDS data on all patients and are required 
to collect and report such data on all patients whose care is paid for under 
federal programs. 

The UHDDS was initially identified as the minimum set of required information 
PSRO's would supply on each federal patient to the federal government. How- 
ever, since information on the review process itself was lacking, a set of 
PSRO specific data elements were added (Part B) to provide this information. 

The PHDDS therefore, has 2 parts: 

PART A - UHDDS 

PART B - DHEW/HSQB PSRO Specified Data I terns 

In those areas with functioning PSRO's, hospitals are required to colisct 
and report the PHDDS for all patients whose care is paid for under federal 
programs. The PSRO in turn, is required to provide HSQB (formerly BQA) 
with data tapes containing the PHDDS on all federal patients In the PSRO 
area after having deleted physician and patient identifiers. 

Many PSROs collect data elements other than PHDDS for local i:se. These 
data elements would also be available for profiles developed at the local 
1 eve 1 . 



•I. 

2. 



6. 
7. 
8. 

*9. 

'10. 

11. 
t*. 

13. 
I*. 



PSRO HOSP ITAL D I SCH ARG E DATA S ET ( PM I DDS^ 
PART A 

Unlforu Hospital Discharge Date Set (UH&6S) 



Person Identification 

Bate of Birth 

Sex 

■ 

Resld'wco 

Hospital Identification 
Admission Rite and Hour 

1 'srt ii :,- lote 

Attending Physl-.laf. Identl Flcation 
Operating Physician Identlf I cat I on 
Diagnosis 

Procerfni-ffS Perforraad 

Ols|x>sltlcrr of Patient 

Exoectel Principal Source of Payisent 



OHF./MSOJ 



«WT e 

.••SHO Sr-elf !c Uata Itaast 



IS. 
* 

1*. 
19. 
20. 

'«*»; 

12. 
1 eVJfS; 



Hmtxsr of Days Certified et AAaisslcn 
Admission Ceri If icatlen ?ror»a« ?-»d Outc 
B»s.ls for Asalmwnt of IrMtlal l^weth-r.f-'itjrv 
Admission Certification level of Ar-vlrv 
Total IMabw 0«vs Ortlf i~i 

Total Hupber of Iwictn Keferras? to Pfyslel"! '-Svlanr 
Total Ntaaher of Uttamsierfs Approv-rd 
Extension Danlals 



V 



These dotn eloiumte cere aolleatod n<j r!vii>w<! by 
the PSRO. Jfiay crre. not. tmtr~itt*H to PSgS/l'vHr', 



18 



In addition, much data has already been accumulated through performance of 
medical care evaluation studies in individual hospitals 



HOSPITAL MCE STUDY^f 



Topic: Primary C-Section 

Criteria: Blood Transfusion 0%, Exception - NONE 



10 



* Variations 

20 



30 



40 



19 
13 

21 
14 
05 
03 
26 
32 
17 
28 
18 
25 
15 
08 
10 
31 
27 
30 
04 
01 
29 
23 
06 

11 
20 
02 
12 
24 
09 
22 



50 



60 



*********** ************************************************ ********* 

HtMHHMHtWHHMtHmHMHHMIUHUHHMUHHU 
*********** *************** 
************************* ************** 



************************************ 
****************************** 
****************************** 
****************************** 
***************************** 
************************ 
************************ 
************************ 



******************** 

****************** 

**************** 

************* 

************ 

********** 



********* 

******* 

**** 



PSRO AREA-WIDE MCE 




Topic: Elective Primary C-Section -Calendar Year 
Criteria: Low Cervical Approach 100%, Exception: 



1977 

Concomitant Sterilizati 
Placenta praevia 



HOSPITAL 






X Variations 


i 


SIZE 
(BEDS) 


H0SP.T0T. 
* PTS. 


| 

3 20 40 60 80 100 1 




11 


3 




1 

1 


LESS 
THAN 


8 

13 


7 
4 
3 






50 


14 
20 


5 








21 


5 


*********************************************************** 


1 




26 


2 




j 




2 


11 




j 


50- 
99 


10 
12 
17 


14 
10 
9 


■ 


1 




18 
24 


16 








.1 








25 


8 








27 


6 








4 . 


19 








S 


47 






100- 

249 


9 

19 


39 
20 














22 
30 
36 




***************************** 






30 

48 




1 




41 


22 




! 




1 


47 


i 


1 




3 


62 


** 




250+ 


7 
15 
23 
29 
31 
38 


51 
48 

39 ' 
44 
57 
40 


i 

*********** 

** 


I 

i 
•I 

1 

1 
i 



The issue of reliability of data will be discussed later. 



19 



REFERENCE POINTS 



REFERENCE POINTS are used as benchmarks for comparison to MEASUREMENTS of 
current performance. For such comparisons to be valid, REFERENCE POINT VAL- 
UES must be of the same ELEMENTS and in the same UNITS as the MEASUREMENTS. 



PUBLISHED NORMS ~ from national or regional data 



PSRO -B 

SERIES 2 

TABLE 2 

PAGE 3 



5FLECTFD DETAILED 
DIAGNOSIS GROUP 
( 1 > 



STAY PROFILE FOR 18 SELECTED DIAGNOSIS GPO/lPS 
NATION. NOpTHEAS T ERN BEGTON. AND PSRO 

JANUARY - JUNE 197/ 



KEY: NATION ♦ N ♦ 

REGION •— R « 

PSRO + P ♦ 

(25TH TO 75TH PERCENTILE. FIRST AMD LAST ♦ 
MARK END POINTS. LETTER MARKS MEDIAN) 



-ALL TITLES- 
TOTAL - f— DAYS STAY-2 

PATIENT5-1 I 2 5 A 5 6 I 8 9 11/ 11 12 13 14 15 16.17 1.8 19 20 21 22 23 24 25 >25 
<2> 



D1S OF GB > BILE DUCT WITH OP.U/0 2ND DX 

NATION 4006 

REGION 1565 

PSSD 86 

DIS PROSUTE. W/TRANSUR PROST. WO 2ND DX 

NATION 4674 

RFMON 130<i 

PUD 54 

DIS OF FEM GEN ORG U/HY5T ER EC . OT H MA J OP 

NATION 7770 

REGION 2708 

PSRO 243 

NORMAL DELIVERY. U/O OP OR W/MINOR OP 

NATION 29704 

REGION 9694 

PSRO 1626 

ARTHRITIS U/REP AIR. PLASTIC OP . LAMNECTOMY 

NATION 4J27 

REGION 1046 

PSRO 43 

FRACTURE OF FEMUR. PELVIS. MULTIPLE W/O OP 

NATION 5310 

REGION 1469 

PSRO 127 



»— N— « 
*--R* 
»P--» 



1- EXCLUDE5 DEATHS 

2- IXCLUDES DEATHS. IF 75TH PERCENTILE EXCEEDS 25 DAYS. IT WILL BE SHOWN AS >25 DAYS. 



HISTORICAL DATA for comparisons over time would be obtained by generating 
this profile for a different time period. 



■ 



20 



GROUP VALUE - for comparison to SUB-GROUPS 
(or SUB-GROUP to SUB~GROUP) 



SERIES 3 
TABLE 4 
PAGE 1 




123 
PATIENTS 
AVG PRE-OP 

210 
PATIENTS 
AVG PRE-OP 

234 
PATIENTS 
AVG PRE-OP 

567 
PATIENTS 
AVG PRE-OP 

654 
PATIENTS 
AVG PRE-OP 

765 
PATIENTS 
AVG PRE-OP 

78? 
PATIENTS 
AVG PRE-OP 

876 
PATIENTS 
AVG PRE-OP 



26 

3.0 



23 
4. 1 



23 
9 8 



75 
3 3 



20 

1 , 1 



2 

A . 



3 

4 . 7 



10 

3.3 



45 

1.2 



24 
1.5 



16 

3.3 



20 
1 A 

20 
2.6 



61 
1 . 1 



19 
1 . 9 



120 

1 2 



37 
2 4 



32 
1.8 



1 

1 . 



19 

1.8 



5 

2. A 



20 
2 . 9 



5 

4 . 



3 

' 



10 

11. « 



7 

9.3 



13 
3.3 



32 

5.7 



25 
S3 



92 

A. 1 





c 



» 1 



15 
A . 9 



11 
5.7 



20 

6 . 7 





. 



1A8 

. 3 



2 

a . 5 





. 



7 

l.A 



11 
2.0 



67 
1.9 





. 



73 
1.0 





. 



301 

1.2 



1 

A.O 



5 



2 



2.3 
A 

3.0 
II 



19 

2.2 









2 

2.5 



2 

! 



18 

5.1 



S 

2.3 



1A 
5.1 





0.0 



112 
0.9 





0.0 



35 
0.6 





. 



24 

l.A 



8 

0.3 





. 



>EtCIUrfS PATIENTS GROUPED UNDER ' 'OTHER DIAGNOSIS GROUPS ' • IN SERIES 3 TABLES 1 AND ? 
PROCEDURE HEADINGS REFER TO THE SAME "ROCEOURES DESCRIBED WITHIN SERIES 3 TABLES 1 A 



* CRITERIA and STANDARD - from author i tative statements of ideal 
performance used in medical care evaluation studies / 





HOSPITAL MCE STUDY / 


Topic: Primary C-Section ./ 


Criteria: 


Blood Transfusion 0%, Exception - NONeI^ 




- 

i Variations 


presicua 




NUMBER C 


10 20 30 40 50 60 


19 


*»****«**««..*.««*«««*««««*••***«***«*****»»*•**««.**«**..»«*«»»«*» 


13 


»««««*«*»«* «««**^*»«»«««»«*««**«*>>t>>»*>**»*>»* tt ** ttt ** t * t 


21 


mHMH>UHMUIIHMM»flMtt*lUMHUimi 


14 


*************************************** 


05 


************* ************************** 


03 


************************************ 


26 


****************************** 


32 


****************************** 


17 


****************************** 


28 


***************************** 


18 


************************ 


25 


************************ 



Thresholds which are pre-defined levels of performance below which some action 
is deemed mandatory are developed from statements of achievable and acceptable 
levels of care reflecting the current state of the art. 



21 



EXAMPLE 



THE GROUP: 

SUB-GROUPS: 
MEASUREMENT: 

REFERENCE POINT: 



All Patients (TYPE) in the PSRO area (UNIVERSE) 
who underwent a primary C-Section (CHARACTERISTIC) 
during January-June, 1977 (TIME PERIOD) 

Hospitals listed in order of bed-size 

The percentage (UNITS) of variation from the 
established criteria (ELEMENT) 

An MCE criterion that states that a low cervical 
approach should always be used in a primary C-Sec- 
tion unless the patient has placenta praevia or a 
concomitant sterilization is performed. 




GROUP 

REFERENCE PO 
SUB-GROUPS 



PSRO AREA-WIDE MCE 



Topic: Elective primary C-Section -Calendar Year 1977 

iteria: Low Cervical Approach 100$, Exception: Concomitant Sterilization or 









Placenta praevia 


HOSPITAL 






% Variations 


SIZE 
(BEDS) 


H0SP.T0T. 







PTS. 


3 . 20 40 60 80 1C 




11 


3 




LESS 
THAK 

.50 


8 
13 
14 


7 
4 
3 


*r** ************ ********** ** ***** ******* 


20 


5 






21 


5 


a********************************************************** 




26 


2 






. 2 


11 




50- 
99 


10 
12 
17 
18 
24 
25 
27 


14 

[W 
9 

16 

[_5 
1 8 

r 6 






4 . 


19 






5 


47 




100- 


' 9 


39 




249 


19 


20 


******************************************* 




22 


38 


***************************** 




30 


30 


***************************************************** 




36 


48 






41 


22 






1 


47 






3 


62 


** 


250+ 


7 

15 


51 
46 


*********** 




23 


39 






29 


44 ' 






31 


57 


** 




38 


40 : 





22 



GROUP- 



SUB-GROUPS- 



MEASUREMENTS- 



THE GROUP : 

SUB-GROUPS: 

MEASUREMENTS: 

COLUMN NUMBER 
2 

3 
h 

5 
6 
7 
8 

9 - 13 
REFERENCE POINTS - 



EXAMPLE 



Patients (TYPE) in PSRO Area X (UNIVERSE) in 
selected diagnosis groups (CHARACTERISTIC) 
during Jan-June, 1977 (TIME PERIOD) 

18 selected diagnoses listed in diagnosis 
index code numerical order (not printed) 



(8 different measurements) 

ELEMENT 

Occurrence of 
each diagnosis 

Days of care used 
by patients 

Occurrence of 
death 

Long stay pateints 

Studied patients 

LOS 

LOS 

LOS 



UNITS 

Total number of 
pat ients 

Total number of 
days 

Total number of 
deaths 

Total number 
Total number 
Average stay 

Coefficient of Variation 
Percent i les 



Reference points do not necessarily have to be in 

the same report, i.e., another report may be required. 



UTILIZATION BY -.FIF.C'Fn PEM'LED DIAGNOSIS CROUPS 
" Pi SC. B 

UAUOMif JUNE 197? 



SERIES 2 
TABLE 1 



-ALL TITLE DISCHARGES- 



SCIECTED DETAILED DIAGNOSIS GROUP 
( 1 ) 

DI A3ETES . AGE -35 U/0 OP U/0 2ND DX 
NEUROSES. PERSONALITY DISORDERS 
DIS OF EYE WITH EXTRACTION OF LENS 
ACUTE MYOCARDIAL INFARCTION 
ISCHEMIC HEART DI5.U/0 OP U/0 2ND DX 
CEREFJ THROMSO- EMBOL U/0 OP. U/0 2HD DX 
HYPERTROPHY OF T t A 

ACUTE UPPER RCSP. INF.JINFLU. AGE>44 
PNEUMONIA, AGE-31 

PNEUrtONIA. AGE*30.U/O OP W/0 2ND DX 
GASTRIC i PEPTIC ULCER U/0 OP U/0 2ND DX 
HERNIA OF ABD CAV. AGE>64 U/MINOR REPAIR 
DIS OF GB > BILE DUCT UITH OP.H/O 2ND DX 
DIS PROSTATE .U/TRANSUR PROST.W/O 2ND DX 
DIS OF F CM GEN ORG U/HYST EREC . OTH MAJ OP 
NORMAL DELIVERY. U/0 OP OR U/MINOR OP 
ARTHRITIS U/ REP A I R. PLASTIC OP . L AMNEC TOMY 
FRACTURE OF rEMUR. PELVIS. MULTIPLE U/0 OP 



-AVERAGE STAY- 



PERCENTILE STAY IN DAYS-' 



TOTAL NO. OF 
DISCHARGES 
(2) 


T01AL 

DAYS 
(!) 


DEATHS 
(4) 


LONG STAY 
FTS. -1 
15) 


STUDIED 
PTS.-2 
161 


AVG. 
STAY 
[ 7 ) 


COEFF 
VAR-3 
(8) 


10TH 
(9) 


25TH 
(10) 


50TH 
(11) 


75TH 
(12) 


90TH 
(13) 


IS* 


1 95 3 


6 


13 


135 


10.5 


0.5 


4 


7 


10 


15 


27 


74 


1362 





1 


73 


17.6 


. 7 


4 


8 


16 


24 


3' 


576 


3702 


2 


23 


551 


5.8 


0.4 


4 


5 


5 


7 


15 


582 


10061 


109 


17 


462 


17.2 


0.5 




10 


16 


24 


33 


119 


1155 


1 


1 


110 


7.8 


6 


3 


5 


8 


10 


20 


101 


11130 


15 


9 


77 


13.4 


. 5 


5 


7 


14 


20 


M 


255 


432 





4 


251 


1 .6 


0.4 


1 


1 




2 


2 


21 


295 





1 




! 6 


0.7 


3 


4 




13 


24 


210 


1 345 


3 


6 


201 


5.8 


0.5 


J 


4 


5 


5 


11 


42 


480 


5 


2 


35 


8.8 


0.5 


3 


5 


9 


13 


19 


35 


270 


1 


1 


33 


7.1 


0.4 


3 


5 


7 


» 


12 


252 


247 1 




12 


238 


8 . 5 


. 5 


4 


6 


8 


11 


1? 


86 


1 146 


a 


6 


8 


11.9 


0.4 


7 


8 


11 


16 


23 


54 


621 








54 


11.5 


0.4 


6 


9 


11 


14 


16 


244 


2106 


l 


? 


236 


7.8 


5 


4 


5 


7 


10 


14 


1626 


5657 





19 


160 7 


3.4 


0.3 


3 


3 


3 


4 


5 


43 


1565 





7 


36 


29.5 


. 3 


16 


25 


30 


♦ 3 


69 


132 


3644 


s 


13 


114 


21.8 


0.7 


4 


11 


21 




41 



1- LIVE DISCHARGES UITH UNUSUALLY LONG STAYS. WHO UOUl D 

DEPENDENT UPON THE PATIENT'S PRINCIPAL DIAGNOSIS. 

2- E-:CLUDES MATHS AND LONG STAISICOLS. 4 A MB 5) THESE 

3- THE STANDARD DEVIATION DIVIDED BY THE AVERAGE STAY 

4- EXCLUDES DEATHS ONLY. 



SKEW THE AVERAGE STAY. THE 
PATIENTS ARE ALSO EXCLUDED 



DEFINITION OF A LONG STAY IS 
FROM COLS. 7 AND 8. 



23 



WHAT NEW INFORMATION COMES FROM 
A PSRO PROFILE 



DESCRIPTIONS 

PSRO profiles provide a description of various dimensions of a patient/ 
provider/institutional group such as: 



utilisation by selected procedures 
january - june 19;/ 



PSRO 3 



SERIES 3 
TABLE 1 
PAGE 1 



TOTAL TOTAL DAYS 

DISCHARGES 

t 2 1 ( 3 ) 



PROCEDURE 
( 1 ) 

CHOLECYSTECTOMY 
DIS OF CP- > BILE DUCT U/0 DX2 
DTS or .3 I DUE DUCT U/DX2.AGE'65 
HIS 111 Gil « (III E DUCT U/DX2.AGE 44 
'.L P. TOTAL 

OTHER DIAGNOSIS CROWS 
1UUL DIAGNOSIS GROUPS 

HEMureiioiorcToriY 

HEMORRHOIDS 

I HER DIAGNOSIS GROUPS 
TOTAL DIAGNOSIS GROUPS 

INGUINOFEMORAL HERNIORRHAPHY 
HERNIA OF ADD CAV, ACC<15 
l.'IG HERNIA Ll/O OBS. 14^AGE<65U/0 DX2 
HERNIA OF ADD CAV, AGE -64 
SUBIOTAL 

OTHER DIAGNOSIS GROUPS 
TOTAL DIAGNOSIS GROUPS 

HYSTERECTOMY 
CA UTERUS -CORPUS W/ASD. HYSTERECT. 
CA UTERUS-CERVIX OVARY U/ABDHYSTR 
BEN NEOPL UTERUS. OVARY 
DIS FEMALt CEN. ORGANS 
SUBTOTAL 

OTHER DIAGNOSIS GROUPS 
TOTAL DIAGNOSIS Gk'OUPS 

TRANSURETHRAL PROSTATECTOMY 

CI MALE GENITAL ORGANS 

DIS OF PRO: TATE U'O DX2 

. DIS UF r^OSIATE WDX2 
SUBTOTAL 

OtHEP DIAGNOSIS CPnilPS 

TOTAL DIAGN'JSIS GROUPS 



l-i I v c Disr.iivpcrs with unusuaily iong stays, who would 

DEPENDENT UPON THE PATPNT'S PRINCIPAL DIAGNOSIS 

I-Exciurrs deaths two echo stays <cois. 5 and 6). these pa 

3-IHr S T .".r--D D I V 1 1 T I '. N DIVIDED BY THE AVERAGE STAY. 



AVG. FRE- 
ST AY 
(A) 



LONG STAY 
FT5 . -I 
(61 



AVERAGE STAY 

STUDIED AVG . COEFF. 
'IS.-2 STAY VAR-S 

( 7 ) (8> 



( 9 ) 



74 


956 


3 1 





5 


69 


11.4 


. 4 


68 


S14 


3.0 





1 


67 


11.6 


. 4 


iai 


2124 


6 8 


7 


5 


89 


is. 2 


0.4 


:<. 1 


3S94 


4.6 


7 


1 1 


225 


14.1 


. 5 


20 


604 


11.6 





3 


17 


25.2 


0.6 


263 


4 4 98 


5.1 


7 


14 


242 


14.9 


0.6 


61 


464 


2.1 





2 


59 


7.2 


3 S 


5 


67 


4.6 








5 


13.4 


. 7 


6 6 


531 


2.4 







64 


7 . 7 


. 6 


112 


263 


0.8 





1 


109 


2.2 


0.6 


34 


202 


1 . 1 








34 


5.9 


0.3 


221 


2224 


1 . 9 




1 1 


216 


8.3 


. 5 


375 


2694 


1 . 5 




14 


359 


6.2 


0.7 


53 


799 


4 . 5 


3 


3 


47 


11.4 


. 9 




3493 


1 . 9 


5 


17 


406 


6.8 


. 8 


43 


488 


2.1 








43 


11.3 


. 3 


!} 


210 


3 4 








11 


19.1 


. 6 




872 


2.3 





5 




9.9 


. 3 


46 


490 


2.3 





1 


45 


10 2 


. 3 


179 


2060 


2.3 





6 


173 


11.0 


. 4 


25 


345 


3.3 


1 


2 


22 


10.4 


0.4 


204 


2405 




1 


8 


1 is 


10.9 


0.4 


37 


825 


7. 1 


1 


4 


32 


17.2 


5 


45 


506 


3. 1 








45 


11.2 


. 3 


125 


2052 


6 . 3 


3 


4 


lis 


15.1 


0.5 


207 


3383 


S 1 


4 


8 


1 95 


14.5 


5 




Col umn 
(2) 
(3) 
W 
(5) 
(6) 
(7, 8, 9) 



How many discharges are in each group 
How many hospital days did each group consume^ 
What was the average pre-op stay — 
How many deaths were there 



How many long stay patients were there 



What was the average time spent in the hospital 



24 



• DIFFERENCES 



By displaying equivalent values (same elements and units) for group(s) in 
conjunction with reference points, differences and similarities become 
readi ly apparent. 



PSRO -B 

5E3I5S 2 
TAr.LE 2 
PAGE 5 



SELECTED DETAILED 
DIAGNOSIS GROUP 
(1) 



KEY I NATION » N ♦ 

UK PROFILE FOR It SELECTED DIAGNOSIS GPm'»S REGION »—»-—* 

NATION. NfipTHfASTfRM »I5!0N. AND PSRO R PSRO ♦ P • 

(25TH TO 75TN PERCENTILE. FIRST AND LAST « 

JANUARY - JUNE 1»7' HARK END POINTS. LETTER HARKS MEDIAN ) 

-All TIKES- 

TOTAl DAYS STAY2 

PATIENTS-) 1 I ] ♦, J t 7 • 9 10 11 12 13 1 A IS 16 17 18 1 9 21 21 22 23 2'. 25 >2S 



DIS OF GB A BILE DUCT WITH 0P.U/O 2ND DX 

NATION AOOt 

REGION 15»S 

PSRD 86 

DIS PROSTATE. U/TRAMSUR PROST. U'O 2ND D« 

NATION A»?« 

REGION UK 

PSRO 5* 

DIS OF FEB GEN ORG U/HY5T EREC . OTM MA J OP 

NATION 7770 

REGION 2J0« 

PSRO JAJ 

NORMAL DELIVERY. UVO OP OR U/M1NOR OP 

NATION 2»7|* 

REGION 9S9A 

FSRO l»2i 

ARTHRITIS U/REPAIR.PIA5TIC OP . L AMNECTOMY 

NATION A327 

REGION 1C» 

PSRO «1 

FRACTURE OF FEMUR . PELVIS. MULTIPLE UVO OP 

NATION 5 510 

REGION 1469 

PSRO 127 



«--N • 



•--N--« 



1- EXCLUDES DEATHS . 

2- EXCLUDES DEATHS. IF 75IH PERCENTILE EXCEEDS 25 DAYS. II Will BE SHOWN AS >25 DAYS. 



FOR EXAMPLE: 

• For Disease of GB and Bile Duct the median stay for the PSRO is 11 
days, while that for the region and nation is 10 days. 



25 



WHEN IS A DIFFERENCE MEANINGFUL? 



There are many sophisticated statistical methods for testing the hypothesis 
that different values represent a real difference. But for practical pur- 
poses a difference is meaningful, i.e., leads to further activity, when it 

i s big or repeated. 



BIG 



Two values are clearly separated by a large difference. 



EXAMPLES: 

Hospitals #14 and 
#21 clearly have a 
higher % of varia- 
tions than the 
others in the less 
than 50 bed size 
group. For hos- 
pital #21 the 1003 
variation rate 
represents 5 cases, 
all of which did 
not comply with 
the criterion. 



The 250+ bed size 
group clearly has 
fewer significant 
variations than 
the 100-2^9 bed 
size group. 











PSR0 AREA-WIDE MCE 


Topic: Elective Primary C-Section -Calendar Year 1977 


Criteria: Low Cervical Approach 1003, Exception: Concomitant Sterilization or 








Placenta praevia 










HOSPITAL 






t Variations 


SIZE 


BOSP.TOT. 




(BEDS) 


I 


PTS . 


1 20 40 60 80 100 






11 


3 










8 


7 








LESS 


13 


4 


■ 






THAN 










50 : 


r 14 


3 


****** ********************************* 






20 


5 










m 


5 


*********************************************************** 








26 


y 










2 


ii 


t 






50- 


10 


14 








12 










99 


10 










17 


9 










18 


16 










24 


.5 


************************************************* 








25 


8 










27 


6 










4 . 


19 










5 


47 








100- 


9 


39 


■ 






249 


±9 


20 


******************************************* * 








22 


38 


***************************** t 








30 


30 


***************************************************** 








36 


48 










41 


22 










1 


47" 










3 


62 


** 








7 


51 








250+ 


15 


48 


*********** 








23 


39 1 










29 


44 : 










31 


57 


** 








38 


40 ' 

















26 



REPEATED 

Multiple measurements all show a difference in the same direction. 



PSRO -B 

SERIES 2 

TAME 2 

PAGE 3 



STAY PROFILE FOR 18 SELECTED DIAGNOSIS GROUPS 
NATION. NflRTftAS'ERN REGION. AND PSRO j3 



JANUARY - JUNE 197/ 



KEY ' NATION ♦ N ♦ 

REGION * R ♦ 

PSRO * P ♦ 

(25TH TO 75TH PERCENTILE. FIRST AND LAST 
MARK END POINTS. LETTER MARKS MEDIAN) 



SELECTED DETAILED 
DIAGNOSIS GROUP 
( 1 ) 



-ALL TITLES- 
TOTAL 
PATIENTS-1 

12) 



DAYS 5TAY-2 

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 >25 



DIS OF GB I BILF DUCT WITH OP.W/0 2ND OX 
NATION 
REGION 

PSRO 

DIS PROSIATE. W/TRAN5UR PROST. WO 2ND DX 
NATION 
REGION 
PSRO 

DIS DF F EM GEN ORG W/HYSTEREC.OTH MA J OP 
NATION 
REGION 
PSRO 



NORMAL DELIVERY, 



W/O OP OR W'MINOR OP 
NATION 
REGION 
PSRO 

ARTHRITIS W/REPAIR. PLASTIC OP . I AMNECT DMY 
NATION 
REGION 
PSRO 

FRACTURE OF FEMUR. PEL VIS. MULTIPLE U/O OP 
NATION 
REGION 
PSRO 



1565 
86 



A67A 
1 304 
54 



77 70 
2708 
243 



29704 
96 94 
1626 



4327 
1046 
43 



5310 
1469 
127 



t--N • 



*--RA 
*p--t 



1- EXCIUDE5 DEATHS. 

2- EXCLUDES DEATHS. IF 75TH PERCENTILE EXCEEDS 25 DAYS. IT WILL IE SHOWN AS >25 DAYS. 



EXAMPLE: For diseases of prostate 

1. The 25th percentile for the PSRO is 2 days 
higher than for the region. 

2. The median for the PSRO is 3 days higher than 
for the region. 

3. The 75th percentile for the PSRO is 3 days 
higher than for the region. 



27 



EXPLAINING DIFFERENCES 

Where meaningful differences exist, they may sometimes be explained by 
obtaining rearranged or additional data such as 

• SUB-SETS 

Dividing a measurement into its component measurements, e.g., dividing 
length of stay into pre-op and post-op stay may often help to focus in 
on the reason for a difference. 



UTILIZATION By SFLCCIED PROCEDURES 
JANUARY - JUNE 197/ 



PSRO 5 



SERIES J 
TABLE i 
PAGE 1 



CMULEf Y5TECI0MY 



OP OB 
DIS OF oB 
DIS OF C.b 
SUBTOTAL 



DUE 0'JCI U'O DX2 
DUE DUCT U/DX2,ACE'.65 
BIl[ DUCT W/DX2.AGE.-64 



TOTAL 
DISCHARGES 
<2> 



68 
101 
243 



OtlltR DIAGNOSIS CROUPS 
TulAl DIAGNOSIS GROUPS 

IIE"OrPUOIDECTOMY 
HI MORRIIOIBS 

OTHER DIAGNOSIS CROUPS 
TOTAL DIAGNOSIS GROUPS 

INGUINOFEMORAL HERNIORRHAPHY 
HERNIA OF ADD CAV, AGE<15 
I !<G . HERNIA U/O OB5 . 1 4< AGE<6 5U/0 DX2 
HERNIA OF ADD CAW. AGE -64 
SUBTOTAL 

OTHER DIAGNOSIS GROUPS 
TOTAL DIAGNOSIS GROUPS 

HYSTEREC TOMY 
CA UTERUS-CORPUS U/ABD. HYSTERECT. 
CA UTERUS-CERVIX OVARY U/ABD. HYSTR 
BTN NEOPL UTEilUS. OVARY 
DIS FEMALE GEN . ORGANS 
SUBTOTAL 

OTHER DIAGNOSIS GROUPS 
TOTAL DIAGNOSIS GROUPS 

TRANSURETHRAL PROSTATECTOMY 

C\ MALE GENITAL ORGANS 

DIS OF PROSTATE U' DX2 

. DIS OF TROSIATE U/DX2 
SUBTOTAL 

OTHER DIAGNOSIS GRfTUPS 

TOTAL DIAGNOSIS GROUPS 



TOTAL DAYS 



814 
2121 
3S94 



20 
263 



(1 
5 
66 



34 

229 
375 

53 
4 28 



43 
II 
79 
4 6 
175 

25 
204 



37 
45 

39 
246 



604 
4498 



t 6 '. 

6 I 
531 



263 
202 
2224 
2694 

799 
3493 



4S8 
210 
S72 
490 
2060 

345 
2405 



825 

506 
2052 
3383 

1233 
4616 









AVERAGE 


STAY 


. PRE-OP 


DEATHS 


LONG STAY 


STUDIED 


AVG. 


COEFF 


STAY 




PTS.-l 


PIS . -2 


STAY 


VAR-3 


(4 ) 


(5) 


(6) 


(7) 


C3> 


( 9 1 


3. 1 





5 


69 


11 


4 


. 4 


3.0 





1 


67 


11 


6 


0.4 


6 8 


7 


5 


89 


18 


2 


0.4 


4.6 


7 


11 


225 


14 


1 


0.5 


11.6 





3 


17 


25 


2 


0.6 


5.1 


7 


14 


242 


14 


9 


0.6 







2 


51 


7 


2 


5 


4 . 6 








5 


13 


4 


. ' 


2.4 







64 


7 


7 


0.6 


1 . t 





3 


109 


2 


2 


. 6 


1 . 1 








34 


5 


9 


0.3 


i 1 


2 


1 1 


216 


6 


3 


0.5 


] 5 




14 


359 


6 


2 


0.7 


*.5 


3 


3 


47 


11 


4 


c 9 


1 . 9 


5 


17 


406 


6 


> 


0.8 


2.1 








43 


1 \ 


3 


3.3 


3.4 








11 


19 


1 


0.6 


2.3 





5 


74 


9 


9 


. 3 


2 . 3 





1 


45 


10 




. 3 


2 . 3 





6 


173 


11 





. 4 


3.3 


1 


2 


22 


10 


4 


. 4 


2.4 


1 


8 


195 


10 


9 


0.4 


7 . 1 


1 


4 


32 


1 7 


2 


5 


3.1 








45 


11 




0.3 


6.3 


3 


4 


lis 


15 


1 


0.5 


5.8 


4 


8 


195 


14 


5 


0.5 


12.9 


1 




34 


23 


3 


. 6 


6 . 9 


5 


12 


229 


15 


8 


. 6 



1-lIVe DISCHARGES WITH UNUSUALLY LONG STAYS . WHO WOULD SKEW THE AVERAGE MAT. TH!I DEFINITION OF A LONG STAY IS 
DEtEKBEMf Vra^ THF TATIfNT'5 PINCIFSl DIAGNOSIS. 

J-EXCIUCES DEATHS *HB long 5IAYS (COLS. 5 *nd 6). these patients are also excluded from cols, s AND ». 

5-THC ST.'.t.pti'D DEVIATION. DIVIDED BY THE AVERAGE STAY. 



For the procedure transurethral prostatectomy the difference in average 
stay between cancer patients (17-2) and those with diseases of prostate 
(11.2) is almost totally due to a difference in pre-op stay. 



23 



CLUSTERS 



Some differences may be due to just a few pat ients/provi ders/ i nst i tut ions 
in the group. Rearranging the data into different SUB-GROUPINGS (for 
example by hospital) may clarify the differences or help rule out certain 
factors. 































PSRC 3 














AVERAGE 


PRE-OP STAT" FOR 
HOSPITAL 

J A h u AK ' 1W7 - 


SELECTED 
DATA 

_> 

JUNE 


PROCEDURES 
19 7, 








SERIES 

TABLE 

PAGE 


3 
4 
1 




































































AREA 
( 1 ) 


CHOIE 
(21 


HErtOR 
(3) 


INGUIN0 
HERN 
<4> 


HYST 
(5) 


TRANS 
PROS! 
(6) 




SUPRA 
PROS! 

(7) 


T 1 A 

(8) 


INTRA 
LENS EXTR 
(9) 


RAD 

MASTEC 
(10 1 


ARTHRO 
PLASTY 
(11) 


C-SEC 
(12) 


NAT I OH 

NORTHEASTERN REG . 
PSRO 


3. 
4 . 
4 . 


S 

4 
6 


1 . 9 
2.1 
2.3 


1 .5 
1 . 7 
1.5 


2 . 
2 5 
2.3 


3 . 
4 . 
5. 


4 
3 
i 


4. 
4 . 
6. 


2 
9 
6 


. 9 
.8 
.5 


1.2 
1.4 
1.4 


2 
2 


} 
7 
7 


2. 
3. 
1. 


5 
1 

2- 


\. 


9 

1 


HOSPITALS 




































123 
PATIENTS 
AVG PRE-OP 


26 

3. 





20 
1 1 


45 

1 . 2 


37 
2 4 


8 

7 . 







. 





49 
1.0 


11 


1 
4 














0. 





210 
PATIENTS 
AVG PRE-OP 


JJ 

4 . 


1 


2 

4 . 


24 
1 . 5 


32 
1.8 


10 

11. 


4 


5 
8 


4 




0.0 


19 
2.9 


7 
2 


3 


4 





112 

0. 


9 


234 
PATIENTS 
AVG PRE-OP 


7 

4. 


1 


2 

2.0 


16 

3.3 


1 

1 . 


7 
9 . 


3 


6 

3 


5 




. 


67 
1 .9 


5 


8 


2 
2 


5 









567 
PATIENTS 
AVG PRE-OP 


24 


t 


10 

1 . 1 


20 
1.4 


1 9 
1.8 


6 

7 . 


J 


6 
S 


7 


14S 

0.3 




. 


2 
2 





2 
3 





35 



. 


(54 
PATIENTS 
AVG PRE-OP 


5. 


4 


J 

6 . 


20 

2.6 


5 

2.4 


13 

3. 


3 


15 
4 


» 




. 




. 


4 
2 


J 


18 
5 


1 









745 
PATIENTS 
AVG PRE-OP 


18 

5 


» 


3 

4 . 7 


61 

1 . 1 


20 
2.9 


32 

5 . 


7 


8 

10 


5 


2 

.5 


73 
1 .0 


4 
3 





8 

2 


3 


24 
1 


4 


iai 

PATIENTS 
AVG PRE-OP 


23 

». 


1 


3 

5 . 7 


19 
1 .9 


5 

q 


25 
8 . 


J 


11 
5 


7 




0.0 




3 . B 


11 


7 


14 

J 


1 


s 




3 


876 
PATIENTS 
AVG PRE-OP 


75 

3 


J 


10 
3.3 


120 

1.2 


30 
2.8 


92 

4. 


1 


20 
6 


7 


7 

1.4 


301 

1.2 


19 

2 


2 


30 


4 









•EXCLUDES PATIENTS GROUPED 
*« P^OCECURE HEADIHGS REFER 


UNDER "OTHER DIAGHOSIS GROUPS ' ' IH SERIES 
TO THE SAME PROCEDURES DESCRIBED WITHIN 5 


3. TABLES 1 AND 
ERIE5 3. TABLES 


2. 
AND 2. 















For cholecystectomy hospital #789 had an average pre-op stay of 9-8 days, 
more than twice the average for the PSRO. 



• ADDITIONAL INFORMATION 

Very often the real cause of a difference cannot be determined from the 
data alone. Then additional explanatory information must be obtained 
from the source of service. In the PSRO case this means shifting the 
burden of explaining the differences to the individual hospitals since 
only there can external influences (e.g., lack of long term care beds) 
or other modifying factors (e.g., x-ray machine out of order) be elu- 
cidated. Specific critical events occur (e.g., epidemics, floods or 
other disasters) that influence the pattern displayed; change in hos- 
pital staffing patterns or practice patterns of individual physicians 
can affect the data. Seasonal variations in the population served by 
hospitals (e.g., migrant workers in agricultural communities, vacation- 
ers in resort areas) and introduction of new medical technology may be 
additional information that helps to explain differences. 



29 



RESULTS OF PSRO PROFILE ANALYSIS 



As a result of analyzing a profile, and the additional profiles and infor- 
mation developed as a result of the preliminary analysis, the PSRO may 
determine one or more findings. Examples are: 

• Performance across the PSRO area is uniform and acceptable. 

• Performance in one or more hospitals indicates a likely 
problem area. 

' The problem does (or does not) seem to relate to indi- 
vidual practitioners. 

• The problem seems to be of such a nature that bringing 
it to the attention of the hospitals or practitioners 
should suffice to correct it. 

• Further in-depth study (MCE) is needed to identify 
cause and attribution of problems. 

• Immediate corrective action must be instituted to 
prevent further problems. 

From these findings, there will flow recommendations for action - but these 
must take into account the level of confidence in the data. 

There has been much unease about taking action based on PSRO profile data 
because of suspicions concerning such data. In order to estimate the 
credibility of the data (and thereby be confident in taking action), two 
aspects must be examined — the validity and the reliability. 



30 



VALIDITY 



Validity has to do with whether the data presented really measures the 
characteristic purportedly being investigated (or in other words - what 
is really being measured?). For example, counting the number of patients 
receiving Medicare benefits describes the size of the group, but says 
nothing about the appropriateness of the care. 

For example, one cannot directly measure savings in hospital bed-days (since 
you cannot count patients who were not there), but that savings can be inferred 
by combining two measures to give the difference in pre-op stays during two 
different time periods for an elective procedure. If the pre-op stay was 
reduced by 1.5 days and there are 30 patients in the group in the current 
time period, the impact is a savings of kS hospital bed-days (30 x 1.5) for 
the current period. 

The validity of such an inference depends both on the validity of the indi- 
vidual measures and the validity of the logic used to draw the inference. 
Care should be taken to avoid the tendency seen in many instances of reach- 
ing conclusions that are beyond that which validly can be inferred from the 
data presented. 

For example, just because the mortality rate for Mis in an area increases 
after the installation of an emergency medical system mobile intensive care 
vehicle does not mean that such patients are receiving poorer care because 
they are being treated by technicians. It often means that the rescue 
service has been able to bring a significant number of patients who would 
have died outside the hospital into the hospital and only some of them sur- 
vived. 

RELIABILITY 

Reliability of data has to do with whether the same measurement gathered 
by different observers is repeatable, and whether the processing of such 
data repeatedly gives the same results. Data quality control techniques, 
such as reabstract i ng a sample of records, make it possible to quantify 
the reliability of data and assist in improving it. For example, counting 
days of stay is usually a very reliable measurement, whereas less confidence 
exists in the reliability of assigning *»-digit diagnosis codes (and even 
less in choosing the "principal procedure"). 

The reliability of computerized data is also a concern. While it is true 
that a computer does not make mistakes (well, almost never) the people who 
abstract and enter data and program and operate computers are just as fal- 
lible as anyone else. Before placing heavy confidence in such data, ade- 
quate and effective quality control procedures should be in place and 
functioning. Such controls should include periodic reabstract i ng (and also 
recording), verification of correct entry ("twice told true"), checks on 
completeness and internal consistency, periodic re-testing of programs, and 
cross-checks between reports allegedly using the same input. Information 
about data quality control procedures and their effectiveness should be 
available to the user so that accurate estimates can be made of the relia- 
bility of the data . 



31 



CREDIBILITY LEVEL vs. "ACTION 



Part of analyzing a finding is to estimate the credibility level of the 
data, considering both the reliability of the measurements and the validity 
of the measures and inferences. 

But even where the credibility level is relatively low, the profile can be 
useful. As long as the predetermined credibility level is accounted for, 
the results of analysis are valid and can be used as a basis for "action." 
An estimate should be made of the credibility level of the data and the type 
of action matched to it. For example: 



Credibi 1 ity Level 
of Data 


Examples of Appropriate 
Types of Action 


HIGH 


• Focus out of 100% concurrent review 

• Withdraw delegated status 


MEDIUM 


• Require explanations from hospital 
or practitioners 

• Indicate MCE Study to determine 
exact cause of problem 


LOW 


• Request other measurements to help 
conf i rm find ings* 

• Observe performance over an addi- 
tional time period 



*Wheve several measurements all seem to point to the same find- 
ing or raise the same question, the validity of such a finding 
or question is muoh higher than if it were based on the analy- 
sis of only one measurement. 



32 



ACTION AND FOLLOW-UP 



After findings have been identified and categorized, some type of action 
must be designated, as well as a mechanism for assuring that the action 
has been (or is being) completed. When PSRO Board members do not consti- 
tute an action group per se } all actions must be implemented and conducted 
through appropriate hospital channels. Even here, however, the PSRO retains 
responsibility for establishing objectives, suggesting actions, and conduct- 
ing fol low-up act i vi ty to assure that action has been taken and objectives 
ach i eved . 

In that the purpose of profile analysis is to identify correctable problems 
or suggest areas where care can be made even better, effective action is 
critical to the entire process. Effective action requires: 

ESTABLISHING OBJECTIVES 

Specific organizational (PSRO) objectives be defined and stated in 
measurable terms. For example, reducing the pre-op length of stay 
for elective surgical procedures to less than or equal to 2 days, or 
the use of whole blood for transfusion only to those patients where 
immediate volume correction is necessary. 

TRANSMITTING RECOMMENDATIONS 

Corrective action recommendations are directed through appropriate chan- 
nels to the medical staff, administrative and board officials possessing 
requisite authority and responsibility to implement corrective measures 
in each hosp i ta 1 . 



33 



TIMETABLE FOR FOLLOW-UP 



Time periods for action to be accomplished are established, reflecting 
the immediacy of the problems identified in terms of demonstrated con- 
sequence to desirable patient outcomes. 

DEMONSTRATING IMPACT 

Comparative profiles showing changes in the values of appropriate measure- 
ments are used to demonstrate the effectiveness of the PSRO review system. 
For example, where a PSRO has addressed the issue of pre-op stay, as a 
result of profile analysis, a profile of a subsequent time period showing 
the differences serves to quantify the effect on pre-op stay. Thus, the 
impact of review on reducing such stay can be evaluated. 

Follow-up plans are specified to determine and record the effectiveness 
of all actions recommended, especially in respect to achieving the stated 
objectives. For example, it was determined through profile analysis that 
all of the hospitals in the PSRO Area had achieved an average length of 
stay for AMI patients that was 1 day lower than the regional norm. While 
the PSRO board members felt this was certainly acceptable, and said so in 
the congratulatory report sent to the hospitals, they also felt that some 
of the more recently published information on selecting AMI patients for 
early discharge was directly applicable to the practices and patient pop- 
ulations in their hospitals. Therefore they took the action of recommend- 
ing that this information be disseminated to the hospital medical staff 
and set an objective of reducing the average stay of AMI patients by at 
least 1 more day over the next 6 months. 



3* 



MONITORING PROGRESS 



Once action has been initiated, some plan must be established for monitoring 
as a mechanism for follow-up, i.e., re-profiling and reanalysis to determine 
the effectiveness of action, and periodic profile analysis to be sure that 
slippage does not occur. The monitoring must detail: 

1. Specifications for profiles to be generated (GROUP, 
SUB-GROUPS, MEASUREMENTS, REFERENCE POINTS). 

These may be the same as the original or some modi- 
fication suggested by the analysis. 

2. A schedule for producing the profile that allows for 
the action to have had an effect. 



35 



AN EXAMPLE 



The following example of the complete process is taken from 
PSRO Transmittal No. 61, dated January 23, 1978, on PSRO 
Prof i le Analysis. 



A PSRO is interested in examining the length of stay pattern in each of the 
hospitals in the PSRO area for 15 selected diagnosis groups, chosen from the 
leading causes of admission in the area. The purpose is to determine whether 
the patterns are consistent across hospitals or any marked differences exist. 
A previous study had indicated that readmission rates for these diagnoses did 
not vary among the hospitals. A quarterly profile report is designed, dis- 
playing for the area as a whole and for each hospital the total number of 
patients in each of the selected diagnosis groups, their total stay and the 
average stay for all live discharges (Table 1). 



ATTACHMENT 1 



THIS REPORT IS PREPARED FOR THE PSRO AREA AND EACH HOSPITAL 

UTILIZATION BY SELECTED DIAGNOSIS GROUP 
PSRO AREA X 



TA8LE I 



SELECTED DIAGNOSIS GROUP 
(1) 


TOTAL 
PATIENTS 

(2) 


TOTAL 
0AVS 

(3) 


DEATHS 

w 


AVERAGE 
STAY • 

(5J 


25t 


Percentile Stay in Days---* 
h 50th 75th 
(7) (8) 


Diabetes of age > 35 without secondary dx 
or with minor secondary dx 


100 


630 





6.3 


3 


5 


7 


Neuroses (Obsess ive-compulsive-depressive} 


75 


729 





9.7 


4 


6 


10 


Diseases of She eye with extraction of lens 


96 


291 





3.0 


3 


3 


3 


Acute Myocardial infarction 


73 


816 


15 


12.0 


/ 


11 


14 


Ischemic Heart disease, without operation, 
without secondary diagnoses 

Hypertrophy of T A A 


80 
98 


420 
215 


2 



5.1 
2.2 


3 
2 


4 

2 


7 
2 


Acute upper respiiatory infection & 
influenza of age ? 44 


44 


510 





11.6 


5 


7 


11 


Pneumonia of age ^ 30 witnout secondary 
diaynoses A without operation 


52 


555 


1 


10.7 


6 


B 


10 


Gastric & peptic ulcer without operation 
& without secondary dx 


10 


63 





6.1 


3 


3 


12 


Abdominal henna of aye^64 with minor 
repair operation 


65 


■142 





6.8 


5 


6 


8 


Disease of gall Madder i bile duct with 
operation without secondary dx 


88 


1.9/ 





10.2 


6 


9 


12 


Disease of prostate with transurethral 
prostatectomy without secondary dx 


75 


555 





; 4 


5 


6 


8 


Dlsea'.e of female genital organs with 
hysterectomy repair, major operation 


160 


1408 





e e 


6 


7 


9 


Norrial delivery without operation or with 
minor operation assisting delivery 


329 


10bb 





3.3 


2 


2 


3 


Fracture of femur, pelvis, multiple 
without operation 


32 


425 





13.3 


4 


B 


15 



Lxcludes deaths 



Also shown are the 25th, 50th, and 75th percentile stays, 



36 



The latter percentile data are then graphically displayed in a second pro- 
file for each selected diagnosis group (Table 2). It is decided in advance 
that attention will be drawn to hospitals with a 50th percentile stay the 
same or higher than the 75th percentile stay for the area. 



TABLE 2 



THIS REPORT WILL RF PRODUCED FOR EACH DIAGNOSIS GROUP 



STAY PROFILE FOR SELECTED DIAGNOSES 
PSRO X AND HOSPITAL DATA 
January - March 1 Q77 



ni'.easp of female oenital ornans with hysterectomy, 

repair, major operation 



API A 
(1 ' 


TOTAL 
"ATI EfJTS 

(?) 




160 


I, 


15 


i#y |i r 


25 


HiiSP r 


IP. 


I icsp n 


23 


HOSP f 


10 


IKTSP E 


14 


MOSP >' 


20 


M|Hp 11 


26 



+ + + 



DAYS STAY* 

8 ) 10 II 1? 13 



+.._n 1 



-._r f 



1« 15 16 17 18 1<5 20 



♦ .__p , 

+ f 



* n + 



-r- -+ 



f — + — +. 



KEY 



•'ududes deaths. If the 75th percentile exceeds 20 days, 

i I wl 1 1 iin of f the seal e. 



PSRO 
HOSPITAL 



LFTTER 



+-.-P---+ 
+.--A---+ 

25th and 75th percentile 

end points 
50th percentile (median) 



The reports are analyzed, and two hospitals are found to exceed the screen- 
ing criteria for five of the studied diagnosis groups. Since each of the 
groups represents operated patients, it is decided at this time to examine 
the pre-operat i ve stay patterns of each hospital in the area for these five 
groups to see if this might help explain the overall longer length of stay 
patterns in the two designated hospitals. 



A third report format is designed, displaying for the area as a whole and 
for each hospital the total number of patients in each of the five groups 
and their average pre-op stays (Table 3) • 



TABLE i 



THIS REPORT IS PREPARED FOR THE PSRO AREA AND EACH HOSPITAL 

AVERAGE PRE-OP STAY 
SELECTED DIAGNOSIS GROUPS 
PSRO AREA X 
January-March 1977 



SELECTED DIAGNOSIS 
GROUP 
(I) 



TOTAL 
PATIENTS 

(2) 



AVERAGE 
PRE-OP STAY 
(3) 



Hypertrophy of T S A 98 

Abdominal hernia of age 
> 64 with minor repair op 65 

Disease of gal 1 bladder 
and bile duct with operation 88 
without secondary dx 

Disease of prostate with trans- 
urethral prostatectomy without 75 
secondary dx 

Disease of female genital organs 
with hysterectomy, repair, 160 
major operation 



.9 
1.4 
2.2 

1 .6 

1.2 



Again, the data also are graphically displayed for each of the five target 
diagnosis groups (Table k) . 



TABLF 4 



THIS RFPORT IS PRODHCFD FOR FACH OF THE FIVF TARGFT DIAGNOSIS GROI'PS 



AVERAGE PPE-OP STAY 
SELECTED DIAGNOSIS GROUPS 
ALL HOSPITALS, I'SRO X 
January - March 1977 



Uisease of female genital organs with hysterectomy, 
repair, major operation 



HOSPITAL 
( 1 ) 



HOSPITAL A 
HOSPITAL R 
HOSPITAL C 
HOSPITAL D 
HOSPITAL F 
HOSPITAL F 
HOSPITAL G 
HOSPITAL H 



TOTAL 
PATIENTS 

(2) 



IE 
?5 
1R 
23 
19 
14 
20 
26 



< 1 DAY 

(3) 



Average Pr e-Op Stay* 



1.5 



P 



3.5 



4.5 



7.5 



Days 



Rounded to the nearest half day for patients with pre-op st.iys of 1 day or greater. 

This follow-up analysis indicates that average pre-op st*y is consistently 
longer in both of the indicated hospitals for each of the five groups studied. 
This difference accounts for much of the overall difference in length of stay 
seen for these hospitals in the first two sets of reports. 



3ft 



Before pursuing this finding directly with the hospitals involved, the PSRO 
considers whether first to examine the pre-op stay patterns of the individual 
physicians in these two hospitals who treat the five groups of patients in 
question. It is decided that, whereas this might provide additional useful 
information about whether the longer stay patterns are institution-wide or 
can be attributed to individual practitioners, there are only sufficient 
data to examine this question for one of the diagnosis groups (i.e., disease 
of female genital organs with hysterectomy, repair, major operation). A 
profile is prepared for each of the two hospitals, displaying the average 
pre-op stay patterns for the physicians who treat patients in this diagnosis 
group (Table 5). The data indicate that, at least for this group of patients, 
the longer stay pattern is institution-wide. 



TABLE 5 



THIS RFPORT IS PTODUCFO FOR ALL PHYSICIANS TREATING PATIENTS 
WITH THIS PRINCIPAL DIAGNOSIS IN HOSPITALS B AND G 



AVERAGE PRE-OP STAY 
SELECTED PHYSICIANS 
HOSPITAL B 
January - March 1977 



Average Pre-Op 'tay* 



PHYSICIAN 
(1) 


TOTAL 
PATIENTS 
(2) 


<1 Day 

(3) 


1 


1.5 


Z 


2.5 


3 


3.5 


32f 


8 










X 






328 


5 










X 






335 


5 








t 








337 


7 










X 







4.5 



6.5 



7.5 



8+ Days 



•Rounded to the nearest half day for patients with pre-op stays of 1 day or greater. 



The PSRO then checks with all the hospitals and le?rns that th^ two affected 
hospitals do considerably less pre-op testing on an out-patient basis than 
the other hospitals. An educational session is planned for the physicians 
from the two target hospitals who perform the procedures under study and for 
relevant administrative personnel. Physicians with similar specialties ar<J 
invited from other hospitals to discuss the policies they follow for out- 
patient testing. It is agreed by the participants from the target hospitals 
to adopt similar policies to shorten pre-op stay. 



39 



A follow-up study is conducted after the new policies have been in effect 
for three months. The graphic display of pre-op stay for each of the five 
target diagnosis groups is repeated for the new time period (Table 6) to 
examine changes in the two target hospitals and to monitor continuing pat- 
terns in the other hospitals. 

TABU 6 

THIS REPORT IS PRODUCED TOR EACH Of THE riVE TARGET DIAGNOSIS GROUPS 

AVERAGE PRE-OP STAY 
SELECTED DIAGNOSIS '.ROUPS 
ALL HOSPITALS, TSRO X 
July - September 1977 



Disease of female genital organs with hysterectomy, 
repair, major operation 



HOSPITAL 

(1) 

HOSPITAL A 


TOTAL 
PATIENTS 

..ID. .. 

13 


< 1 Day 

. (3J . 

A 


1 


I. 


HOSPITAL P 


26 






B 


HOSPITAL C 


19 




C 




HOSPITAL D 


24 


D 






HOSPITAL E 


18 




C 




HOSPITAL F 


13 






1 


HOSPITAl G 


17 






:, 


HOSPITAL H 

- 


„ 

— 




H 





Average Pre-Op st-iy*_ 



<.5 



1 



4.5 



5.5 



6.5 



7.5 



8* Days 



•Poinded to the nearest half day for patients with pre-up sUy, of 1 day or greater. 



A final profile (Table 7) documents by hospit--' the changes in 
between the two quarters studied. 




THIS REPORT IS PRODUCED FOR EACH HOSPITAL IN THE PSRO AREA 

AVERAGE PRE-OP STAY 
SELECTED DIAGNOSIS GROUPS 
HOSPITAL 8, PSRO X 
July-September 1977 



SLLtCTEIi DIAGNOSIS GROUP 

Lii 



— First Quarter — This Quarter-- 

TOTAL AVERAGE TOTAL AVERAGE 

PATIENTS PRE-OP STAY PATIENTS PRE-OP STAY 



DIFFERENCE 
IN 

PRE-OP STAY 



Hypertrophy of 1 & A 

Abdominal hernia of age 

> 64 with minor repair op 

Disease of gall bladder and bile 
duct with operation without 
secondary dx 

Disease of prostate with 
transurethral prostatectomy 
without secondary dx 

Disease of female genital organs 
with hysterectomy, repair, 
major operation 



12 
8 
10 



2b 



1.5 
2.8 
3.7 

2.6 

2.6 



13 
10 
12 



26 



1.0 

1.9 
2.6 

1 .8 

1.6 



-.9 



-1.1 



-.8 



-1 .0 



AO 



c 



1 



> 



c 




I 
!